Renewable Collector Harmonic Scans: Include Cable Capacitance and Converter Impedance

Before adding a harmonic filter to a renewable collector, scan the frequency-dependent network impedance with the actual cable capacitance, transformer data and converter equivalents. Repeat the scan for credible grid and plant configurations. A single total harmonic distortion value or fundamental-frequency aggregated model cannot show where a topology change shifts a resonance or how an inverter controller interacts with that network.

1. Keep two questions separate

A passive network impedance scan locates resonances that can amplify harmonic emissions or background voltage distortion. A dynamic converter impedance/admittance assessment examines controller interaction and stability. Both are frequency-domain studies, but their model requirements and pass criteria differ. A peak in a passive scan is not, by itself, proof of an unstable inverter controller.

AEMO's harmonic guideline calls for detailed plant reticulation and converter equivalents in its Victorian connection process. AEMO's frequency-scanning initiative concerns interaction assessment and is voluntary. Apply the relevant operator's requirements rather than exporting Australian rules as universal mandates.

2. Preserve the collector topology

Request cable lengths, installation-specific capacitance and resistance, transformer frequency-dependent representation, grounding, filter/capacitor states and external-grid impedance envelopes. Specify measurement ports and sequence definitions. Inverter Norton current sources alone are insufficient where their shunt impedances materially influence resonance.

Obtain converter data across active and reactive operating points, including charging where BESS is present. Record controller and firmware versions. The dual-LV impedance matrix guide gives power-frequency procurement inputs; those do not automatically define high-frequency transformer behaviour.

3. Worked illustrative resonance screen

Assume a simplified per-phase positive-sequence network has effective inductance L = 20 mH and shunt capacitance C = 4 microfarad. These are hypothetical equivalent values referred to one voltage level. Neglect resistance and controller dynamics solely for an initial LC screen.

At 50 Hz the first frequency corresponds to harmonic order 11.25. It is not automatically an exact integer harmonic. More cable shifts this simple resonance downward, while lower inductance shifts it upward. Real networks have damping, several modes and frequency-dependent elements, so these values are neither filter tuning recommendations nor predictions of measured distortion.

4. Turn the screen into a case matrix

Choose a frequency range and resolution that resolve the relevant emissions and suspected modes. Refine around peaks instead of relying on sparse integer-order samples. Keep impedance magnitude and phase, not only a graph screenshot. The chosen range must match the study purpose; a controller-interaction scan at low frequencies is not a substitute for the complete harmonic-emission assessment.

5. Evaluate amplification and mitigation together

For each case, combine emissions, background distortion and impedance using the operator-approved method. At a single frequency, Vh = Zh Ih is a useful linear screening relationship, with compatible complex quantities. It does not justify multiplying unrelated worst-case magnitudes from different topologies or operating points.

Compare filter alternatives against voltage distortion, current duty, damping, losses and contingencies. Confirm that capacitor switching does not move another peak onto a significant emission. Converter control changes need vendor validation and, where interactions remain uncertain, targeted EMT assessment. The inverter fault-protection guide addresses a separate controller-dependent behaviour; one model should not be assumed valid for every study purpose.

6. Specify acceptance evidence

Require the native study model, source data, case list, converter equivalents, scan resolution, complex results and mitigation rationale. Tie the final model to as-built cable lengths, transformer connections and installed settings. IEC 60076-1 frames the general transformer specification; frequency-dependent guarantees need explicit agreement.

Commissioning evidence should include the agreed operating states and measurement locations, instrument configuration, background conditions and comparison with the predicted envelope. A benign reading at one dispatch point does not close an untested contingency. Resolve model/measurement differences before changing filter tuning or controller settings, and keep the revised study under configuration control.

7. Frequently asked questions

Can a 50 Hz equivalent be used unchanged for the scan? Not reliably. Cable and converter detail often determines the result.

Does low THD prove that no resonance exists? No. An unexcited mode may be invisible in that operating measurement.

Will adding capacitance always improve power quality? No. It changes resonance frequencies and reactive duty; study every relevant state.

8. References

• AEMO: harmonic studies guideline
• AEMO: frequency scanning
• IEC: transformer specification scope